A1152
Title: Financial risk forecasting in the data centric era
Authors: Richard Gerlach - University of Sydney (Australia) [presenting]
Chen Liu - The University of Sydney (Australia)
Minh-Ngoc Tran - University of Sydney (Australia)
Chao Wang - The University of Sydney (Australia)
Robert Kohn - University of New South Wales (Australia)
Abstract: How data size, model size, and architectural choice affect the performance of neural networks in financial risk forecasting is investigated. Using volatility forecasts for more than 10,000 stocks, it is found that, within the specifications studied, performance gains from expanding the size and diversity of the training data outweigh those from model scaling or architectural choice. It is further shown that, when neural networks are applied to financial data, global estimation is better aligned with their data-intensive nature and enhances their operational viability.
